Letters: Nonnegative independent component analysis based on minimizing mutual information technique

Authors:
Chun-Hou Zheng;De-Shuang Huang;Zhan-Li Sun;Michael R. Lyu;Tat-Ming Lok
Affiliations:
Intelligent Computing Lab, Institute of Intelligent Machines, Chinese Academy of Sciences, P.O. Box 1130, Hefei, Anhui 230031, China and Department of Automation, University of Science and Technol ...;Intelligent Computing Lab, Institute of Intelligent Machines, Chinese Academy of Sciences, P.O. Box 1130, Hefei, Anhui 230031, China;Intelligent Computing Lab, Institute of Intelligent Machines, Chinese Academy of Sciences, P.O. Box 1130, Hefei, Anhui 230031, China and Department of Automation, University of Science and Technol ...;Computer Science & Engineering Department, The Chinese University of Hong Kong, Shatin, Hong Kong;Information Engineering Department, The Chinese University of Hong Kong, Shatin, Hong Kong
Venue:
Neurocomputing
Year:
2006

Citing 6
Cited 2

Independent component analysis, a new concept?

Signal Processing - Special issue on higher order statistics
An information-maximization approach to blind separation and blind deconvolution

Neural Computation
Linear and nonlinear ICA based on mutual information: the MISEP method

Signal Processing - Special issue on independent components analysis and beyond
Blind separation of positive sources by globally convergent gradient search

Neural Computation
Emergence of Phase- and Shift-Invariant Features by Decomposition of Natural Images into Independent Feature Subspaces

Neural Computation
Algorithms for nonnegative independent component analysis

IEEE Transactions on Neural Networks

On blind source separation using generalized eigenvalues with a new metric

Neurocomputing
Semi-nonnegative independent component analysis: the (3,4)-SENICAexpmethod

LVA/ICA'10 Proceedings of the 9th international conference on Latent variable analysis and signal separation

Quantified Score

Hi-index	0.01

Visualization

Abstract

A novel neural network technique for nonnegative independent component analysis is proposed in this letter. Compared with other algorithms, this method can work efficiently even when the source signals are not well grounded. Moreover, this method is insensitive to the particular underlying distribution of the source data. Experimental results demonstrate the advantages of our approach in achieving satisfactory results regardless of whether the source data are well grounded or not.